Gunjan Gupta

@SteeManThere is some background context of why he posted here. He is having problem with inovato quadra support in our build framework that I added a couple of months back. I asked him to check once on inovato's official image as I wanted to make sure that the onboard wifi is xradio one as I read somewhere that some Tanix TX6 boxes have different onboard wifi. As Quadra is said to be a rebranded TX6, I wanted to confirm the same. Sorry for causing the confusion

@n8henrieWhat was the kernel used in inovato image? If I remember it correctly, they also have a 5.15 and 6.1 version. Can you check their 6.1 version? See if the wifi works there or do you get same error there as well BTW, I have asked the @Tenkawa to test xradio again on inovato quadra. I personally don't have the box. I live in India and can't get the same so all of the testing was done by Tenkawa. He does say that he get better performance on edge kernel. So try creating an image from edge branch as well and give it a try. That image will have 6.6 kernel

A colleague just pointed out to me that you are doing docker in docker. May be just run ./compile.sh PREFER_DOCKER=no to avoid spawning docker instance within docker?

I have built multiple images in the last two days. Haven't faced the same. I will suggest you to check your network. Try using different dns server. Or try using different network if possible.

Could you try by setting disable_fw_kms_setup=1 in config.txt. If its there and already set to one, try setting it to 0

Your box looks exactly like https://linux-sunxi.org/Sunvell_R69. We have a board config file for that tvbox. I am not going to promise how well that is going to work, but if you are willing to try building the image yourself, you should be able to create an image for it. You just have to write that image to a sdcard, insert the same in your tvbox and it should start from the same. Follow the https://docs.armbian.com/Developer-Guide_Build-Preparation/ guide to build your own image. Only change from the guide is, for BOARD variable you need to use sunvell-r69

Your post is lacking a lot of required details and hence its hard to help you. Start by mentioning the image or Armbian version you are using and the kernel version being used. It would also help if you can capture logs using armbianmonitor and share the same.

@Fabrizio68I don't see the board listed on our download page. Which means that we don't support this board. You should ask MKS directly for the same probably using their discord channel as listed on their github page https://github.com/makerbase-mks?tab=repositories&q=ar&type=&language=&sort= We can't help you with this as we never supported this board.

@WarHawk_AVG As Werner said, we don't support this board. But taking an educated guess, I will suggest you blacklist sprdwl_ng module or if its loading through a service or from /etc/modules then remove the same from there. That will lower your load average. You will lose wifi support though as that module is the wifi driver.

@surenz Seems like discord went down. So posting here instead As the overlay you are using seems ok, could you try creating your own custom image using https://github.com/armbian/build. I think there is a kernel patch that might be causing the conflict. Try editing patch/kernel/archive/sunxi-6.1/series.conf file to add a - (hyphen) at the start of the line that contains patches.megous/phy-allwinner-sun4i-usb-Add-support-for-usb_role_switch.patch. Then compile using ./compile.sh BOARD=orangepione BRANCH=current. For detailed instruction on setting up build environment and creating an image, see https://docs.armbian.com/Developer-Guide_Build-Preparation/

For separate ext2 boot partition, I think you can pass BOOTFS_TYPE=ext2 to the compile commandline. If you are ok with writing extensions for armbian build framework or to modify the same, the script that handles partitioning is lib/functions/image/partitioning.sh. There are few hook used functions in there. I am not 100% sure if using those in a custom extension will allow to use LVM for partitioning or not, so you have to try that yourself. Worst case scenario, you have to modify partitioning.sh to add support for LVM.

How is the backup created? Is it simply cloned via dd or something similar?

if you want to do it on the device itself, use the following command update-initramfs -u -k all

In your list of steps, I am not seeing that you ever ran depmod to ensure extraction of updated module deplist and symbols information. Without that kernel will not try to load your module other than you force loading it with insmod. Nothing magical about it, The module most likely would also be present inside initrd file and being loaded from there. You never mentioned that you regenerated initrd, so thats a possible explaination

Its ok. I am already aware that the image was broken. It was already tried by someone else from the community. I will try to fix it later, dealing with some personal stuff right now.

As far as I am aware, there is no upstreamed uwe5622 driver, not even in linux-next tree. So if you are aware of one, please do share. I am aware of a patchset though that was posted to bring some support about 3 years ago IIRC, but it was rejected with a comment that a different group is working on the driver for the same. But no such driver is posted yet. Again do share the upstreamed driver if you are aware of one. We would gladly give it a try.

The wifi and bt configuration you have added is not going to make any difference as in case of allwinner boards, driver simply does not read it. You can see it proven by the "dts node for bt_wake not found" message from your dmesg output above. None of the output there is of any use. Its something that is shown on working devices as well. Enabling uart1 is also not going to make a difference as sprdbt_tty provides bluetooth functionality over mmc/sdio interface. If you want to add bt_wake configuration, you have to add a node with compatible string as "allwinner,sunxi-btlpm" and bt wake information needs to be added in a property called bt_hostwake. Coming back to wifi not working even with the overlay, I am kind of clueless here. Its like the chip is not even powered on and hence is not detected by the driver. Its possible that dcdcc or dcdcd regulators might also need tweaking as done in the Xunlong's repository in this commit to provide required power. But other than that I don't have any clue. I would wait for @ALIGMSTEN to come back on this.

Yeah, its just cherry picked nodes taken from Orange PI Zero2's vendor dts. So I do expect it to work fine. Just I don't have device to test in person.

Could you please try using following dtoverlay along with the extraargs line. Save it to a file with name sun50i-h616-zero2-wifi.dts and then use "armbian-add-overlay sun50i-h616-zero2-wifi.dts" to enable the same /dts-v1/; /plugin/; / { reg_vcc33_wifi: vcc33-wifi { /* Always on 3.3V regulator for WiFi and BT */ compatible = "regulator-fixed"; regulator-name = "vcc33-wifi"; regulator-min-microvolt = <3300000>; regulator-max-microvolt = <3300000>; regulator-always-on; vin-supply = <&reg_vcc5v>; }; reg_vcc_wifi_io: vcc-wifi-io { /* Always on 1.8V/300mA regulator for WiFi and BT IO */ compatible = "regulator-fixed"; regulator-name = "vcc-wifi-io"; regulator-min-microvolt = <1800000>; regulator-max-microvolt = <1800000>; regulator-always-on; vin-supply = <&reg_vcc33_wifi>; }; wifi_pwrseq: wifi-pwrseq { compatible = "mmc-pwrseq-simple"; clocks = <&rtc 1>; clock-names = "osc32k-out"; reset-gpios = <&pio 6 18 1>; /* PG18 */ post-power-on-delay-ms = <200>; }; }; &mmc1 { vmmc-supply = <&reg_vcc33_wifi>; vqmmc-supply = <&reg_vcc_wifi_io>; mmc-pwrseq = <&wifi_pwrseq>; bus-width = <4>; non-removable; mmc-ddr-1_8v; status = "okay"; }; PS: The overlay is tested to compile with "dtc -@ -I dts -O dtb -o sun50i-h616-zero2-wifi.dtb sun50i-h616-zero2-wifi.dts". I can't test if it works as I don't have the device.

for uwe5622, when it comes to allwinner, only device tree specific thing needed is to mark the corresponding mmc node status as ok. Other than that, only available option is for configuring bt_wake pin for bluetooth low power mode. But that is optional. So as long as mmc node is present and status is set to ok, wifi should work find once the corresponding drivers are loaded either manually or via kernel commandline.

In the logs, I can see the driver got loaded with modules_load kernel parameter. But as the logs don't contain a list of network interfaces, its hard for me to tell if it worked for you or not. So could you be more specific whether it fixed your issue or not?

After checking the board config, I see that wifi is only enabled by default for legacy build. Could you please try adding the following in /boot/armbianEnv.txt file extraboardargs=modules_load=uwe5622_bsp_sdio,sprdbt_tty,sprdwl_ng Assuming required device tree node is there, this should get your wifi and bluetooth going on 6.1.53+ or 6.5.3+ kernels

Interesting. Wifi should work on 23.8.3. That release was specifically made to fix uwe5622 driver. The edge kernel you used however is not the one that includes the fix. It should have been something like 6.5.3 or 6.5.4, can't remember the exact version though. Could you please share logs with "armbianmonitor -u" or if you don't have network access use "armbianmonitor -U >logs.txt" and upload the logs manually.

Run "sudo armbian-config" then go to System > Other. There should be 23.8.3 6.5 edge kernel in the list

You didn't mention what was the kernel you used that gave you the issue. The error message is simply stating that your gpu overheated and hence the system is shutting down. May be your dtb needs some adjustment to make proper use of thermal subsystem and frequency scaling for gpu.